Image Enhancement on Deep Learning Algorithm for COVID-19 Lung X-Ray Classification

Abstract

The Novel Coronavirus Disease 2019 (COVID-19) is continuously a phenomenon that continues to study for its development throughout the world because of its international emergency status. The act of testing by clinical laboratory experts is a preventive effort to reduce the increase in cases. However, the number of experts is minimal compared to the cases. So, with deep learning, we need the best model for classifying lung disease variants that the world of health can utilize. This study applies several image enhancement techniques to the convolutional neural network algorithm ResNet50 architecture, which produces gamma correction as the best image improvement technique in this study with an accuracy of 0.986. These techniques also have a reasonably efficient time and a good loss value.